You should have completed the Installation and Authentication sections before proceeding. If you haven’t, please complete them first.

​
How Narada’s API Works

Narada has two main API endpoints:

Remote Dispatch

Sends commands to invoke the Narada web agent in your browser. Returns a request ID for tracking.

All prompts must start with /Operator to invoke the web agent.

Task Status

Checks the status of your request using the request ID - whether it’s pending, successful, or errored.

​
Demo: Narada API in Action

Watch this demonstration of how to use Narada’s remote dispatch API with screenshots and JSON schema:

This video demonstrates the complete workflow from sending a remote dispatch request with screenshot capture enabled and JSON schema validation, through automatic browser execution, to downloading the resulting screenshot files and checking task status via the API.

​
How It Works

  1. Remote Dispatch - You send a prompt request to Narada’s API
  2. Browser Execution - The browser with your Session ID receives and executes the command
  3. Status Tracking - Use the returned request ID to check task status
  4. Results - Get your automation results when the task completes

​
Remote Dispatch Request

The remote dispatch endpoint is a POST request that requires specific headers and request body structure.

​
Constructing Your Request

1

Add Required Headers

Include the API key and content type in your request headers.

Headers
{
  "Content-Type": "application/json",
  "x-api-key": "YOUR_API_KEY"
}

See the Authentication section for instructions on obtaining your API key.

2

Build the Request Body

Create the request body with your prompt and session ID.

Required Fields:

  • prompt: Your command starting with /Operator
  • sessionId: Your unique browser session identifier
Basic Request Body
{
  "prompt": "/Operator search for software engineering jobs and report your findings to me",
  "sessionId": "your-session-id-here"
}

See the Authentication section for instructions on obtaining your Session ID.

3

Optional Advanced Features

Enhance your request with these optional parameters:

clearChat - Clears the chat history when the request is dispatched. Set to true to start each request with a fresh chat.

saveScreenshot - Records the agent’s actions for observability

{
  "saveScreenshot": true
}

When enabled, you’ll receive a download URL for the trace screenshots. See the Screenshots section for more details.

responseFormat - Specify JSON schema for structured output (example below)

{
  "responseFormat": {
    "type": "json_schema",
    "json_schema": {
      "name": "search_count",
      "schema": {
        "type": "object",
        "properties": {
          "count": {
            "type": "integer",
            "description": "Number of search results found"
          }
        }
      }
    }
  }
}

This JSON schema example shows how to structure responses for job search results. To learn how to customize the schema based on your specific data requirements, see the JSON Schema section.

callbackURL - Optional webhook URL for asynchronous responses

{
  "callbackURL": "https://your-webhook-endpoint.com/callback"
}

For detailed webhook implementation, see the Webhooks section.

​
Complete Example Requests

curl -X POST 'https://api.narada.ai/fast/v2/remote-dispatch' \
  -H 'Content-Type: application/json' \
  -H 'x-api-key: YOUR_API_KEY' \
  -d '{
    "prompt": "/Operator search for software engineering jobs in Berkeley and count how many results you find",
    "sessionId": "your-session-id",
    "clearChat": true,
    "saveScreenshot": true,
    "responseFormat": {
      "type": "json_schema",
      "json_schema": {
        "name": "search_count",
        "schema": {
          "type": "object",
          "properties": {
            "count": {
              "type": "integer",
              "description": "Number of search results found"
            }
          }
        }
      }
    }
  }'

​
What the Remote Dispatch Returns

The remote dispatch endpoint returns a request ID that you can use to track your task’s progress:

Response
{
  "requestId": "req_abc123def456"
}

​
Task Status Request

Use the request ID from your remote dispatch call to check the status of your automation task.

​
Making a Status Request

1

Use the Request ID

Take the requestId returned from your remote dispatch request and use it in a GET request to check status.

GET https://api.narada.ai/fast/v2/task-status/{requestId}
2

Include Your API Key

Add your API key to the request headers.

curl -X GET 'https://api.narada.ai/fast/v2/task-status/req_abc123def456' \
  -H 'x-api-key: YOUR_API_KEY'

​
Status Response Types

The task status endpoint will return either:

  • pending - Task is still in progress
  • success - Task completed successfully with results
  • error - Task failed with an error

For automatic status checking, see the Polling guide. For real-time notifications, Webhooks are the recommended approach.

​
Congratulations!

You’ve successfully made your first API request and learned how to check task status.

​
What’s Next?

JSON Schema

Learn how to structure responses with JSON schema validation

Screenshots

Understand how to capture and download agent action traces

Prompt Optimizer

Use Narada’s prompt optimizer to improve your prompts

Webhooks

Set up webhooks for real-time task completion notifications

Polling

Implement polling strategies for checking task status

Python SDK

Use Narada’s Python SDK for streamlined development

​
API References

For detailed parameter information, see: